A TWO-STAGE PRODUCTION MODEL IS USED. THE FIRST STAGE RELATES POLICE INPUTS TO THREE OUTPUTS (CLEARANCES, SERVICE CALLS NOT RELATED TO CRIMES, AND PATROL) PRODUCED ENTIRELY BY THE DEPARTMENT. THE SECOND STAGE RELATES DEPARTMENTALLY PRODUCED OUTPUTS TO SIX FINAL OUTCOMES WHICH POLICE WISH TO MINIMIZE BUT OVER WHICH THEY EXERCISE LIMITED CONTROL. THESE FINAL OUTCOMES ARE TOTAL VICTIMIZATIONS, UNREPORTED PROPERTY VICTIMIZATIONS, COMMUNITY COSTS, LACK OF CONFIDENCE IN THE POLICE, TOTAL PART 1 (RELATIVELY SERIOUS) VICTIMIZATIONS, AND PART 1 PROPERTY VICTIMIZATIONS. THE FIRST STAGE EMPLOYS POLICE DEPARTMENT DATA FOR 81 OF THE CITY'S 245 POLICE BEATS FOR 1971 AND 1972. THE SECOND STAGE EMPLOYS NATIONAL CRIME SURVEY DATA FOR THE SAME YEARS. IN ADDITION TO THE PRODUCTION ANALYSIS, SIX STANDARDS BY WHICH POLICE SERVICES CAN BE DISTRIBUTED IN A COMMUNITY ARE CONSIDERED, AND THE STANDARD THAT BEST DESCRIBES SERVICE DISTRIBUTION IN BALTIMORE CITY IS IDENTIFIED. THE STATISTICAL TECHNIQUE IS ORDINARY LEAST SQUARES REGRESSION. THERE ARE FOUR MAJOR FINDINGS: (1) POLICE ACTIVITIES IN BALTIMORE HAVE A SIGNIFICANT IMPACT (IN THE DESIRED DIRECTION) ON ALL FINAL OUTCOMES INVESTIGATED, (2) RESOURCES USED TO PRODUCE PATROL ARE MORE EFFECTIVE IN BRINGING ABOUT THE DESIRED OUTCOMES THAN ARE RESOURCES USED TO PRODUCE CLEARANCES, (3) FOOT PATROL HAS A GREATER DETERRENT EFFECT THAN VEHICLE PATROL, AND (4) SERVICE DISTRIBUTION IN BALTIMORE IS IN ACCORDANCE WITH AN EFFICIENCY RATHER THAN AN EQUITY STANDARD. THE LIMITATIONS OF THE FINDINGS ARE NOTED, AND IMPLICATIONS FOR POLICY AND OPERATIONS IN THE BALTIMORE DEPARTMENT ARE DISCUSSED. A LITERATURE REVIEW, A BIBLIOGRAPHY, AND SUPPORTING DATA ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--LKM)